Enhancement of Natural Convection by Carbon ... - ACS Publications

Oct 28, 2016 - CNTs would have promising applications in thermal manage- ment devices. Although the thermal conductivity of CNTs have been richly ...
0 downloads 0 Views 5MB Size
Research Article www.acsami.org

Enhancement of Natural Convection by Carbon Nanotube Films Covered Microchannel-Surface for Passive Electronic Cooling Devices Guang Zhang,† Shaohui Jiang,‡ Wei Yao,*,† and Changhong Liu*,‡ †

Energy Conversion Research Center, Qian Xuesen Laboratory of Space Technology, China Academy of Space Technology, Beijing 100094, China ‡ Tsinghua-Foxconn Nanotechnology Research Center and Department of Physics, Tsinghua University, Beijing 100084, China S Supporting Information *

ABSTRACT: Owing to the outstanding properties of thermal conduction, lightweight, and chemical durability, carbon nanotubes (CNTs) have revealed promising applications in thermal management materials. Meanwhile, the increasingly popular portable electronics and the rapid development of space technology need lighter weight, smaller size, and more effective thermal management devices. Here, a novel kind of heat dissipation devices based on the superaligned CNT films and underlying microchannels is proposed, and the heat dissipation properties are measured at the natural condition. Distinctive from previous studies, by combining the advantages of microchannels and CNTs, such a novel heat dissipation device enables superior natural convection heat transfer properties. Our findings prove that the novel CNT-based devices could show an 86.6% larger total natural heat dissipation properties than bare copper plate. Further calculations of the radiation and natural convection heat transfer properties demonstrate that the excellent passive cooling properties of these CNT-based devices are primarily caused by the reinforcement of the natural convection heat transfer properties. Furthermore, the heat dissipation mechanisms are briefly discussed, and we propose that the very high heat transfer coefficients and the porous structures of superaligned CNT films play critical roles in reinforcing the natural convection. The novel CNT-based heat dissipation devices also have advantages of energysaving, free-noise, and without additional accessories. So we believe that the CNT-based heat dissipation devices would replace the traditional metal-finned heat dissipation devices and have promising applications in electronic devices, such as photovoltaic devices, portable electronic devices, and electronic displays. KEYWORDS: carbon nanotubes, heat dissipation, carbon nanotube films, natural convection heat transfer coefficients, thermal management devices



INTRODUCTION Heat conduction and dissipation of hot spots with high localized heat flux, such as portable electronic devices, spacecraft, and photoelectric devices, have become a critical issue of the rapid development of modern science and technology.1,2 The integration levels of electronic devices continue to increase, and the redundant generating heat results in the ongoing increase of working temperatures of related devices. Nowadays, heat dissipation performance of the thermal management devices could seriously influence the working efficiency and lifetime of electronic devices. The key points of developing heat dissipation devices are enhancement of the heat dissipation properties without increasing the volume and the energy consumption of these devices. According to different © 2016 American Chemical Society

working principles, the heat dissipation devices could be divided into two categories, namely, active and passive heat dissipation devices. The active heat dissipation devices, such as the forced convective fins3,4 and thermoelectric heat dissipation devices,5−7 could have perfect heat dissipation properties. Meanwhile, auxiliary accessories and additional electric energy are generally needed in the active heat dissipation devices, which would restrict their applications in portable electronics, special spacecraft, and integrated photovoltaics.8 So the passive heat dissipation devices without other accessories and addiReceived: July 18, 2016 Accepted: October 28, 2016 Published: October 28, 2016 31202

DOI: 10.1021/acsami.6b08815 ACS Appl. Mater. Interfaces 2016, 8, 31202−31211

Research Article

ACS Applied Materials & Interfaces

Figure 1. (a−e) Preparation process of the CNT-based heat dissipation device. The microchannels were etched on copper plates by high-power laser. The SACNT arrays were fabricated by the CVD method, and monolayer SACNT films could be directly drawn from the SACNT arrays by a uniform-speed motor.37 Then SACNT films were spread on the etched copper plates, and ethanol microdrops were dropped by an ultrasonic nebulizer. (f) The optical microscope photograph of etched microchannels on the copper plate and (g) SEM image of lying and suspending monolayer SACNT films. The image shows that the monolayer SACNT films consist of individual CNTs and CNT bundles.

tional electric energy consumption, such as heat pipes9,10 and natural convective fins,8,11 have advantages of small size, freenoise, and energy-saving, which would have promising applications in portable electronics, spacecraft, and photovoltaics. Different kinds of natural convective fins, such as macrofins, microfins, and pin-fins have been considered as effective means to enhance the natural convection heat transfer properties of passive heat dissipation devices. The heat dissipation properties of macrofins have been rigorously investigated so far,12 and the heat dissipation properties of microfins have also been richly investigated during the past several decades. For example, Kim et al.12 have fabricated several microfins with array spacing ranges from 30 to 360 μm on bulk silicon wafers. They found that the enhancement of heat dissipation properties by the microfins over the flat plate was at most 10%. Micheli et al.11 and Mahmoud et al.13 have studied the influence of microfin sizes on natural convection heat transfer properties. Their results showed that the natural convection heat transfer coefficients increase with microfin spaces and decrease with microfin heights.11,13 Although previous works have made great effort to improve the natural convection heat transfer properties of the passive cooling devices, improvements of the natural convection heat transfer property are extremely limited. For example, the best results of the natural convection heat transfer coefficients in traditional microfinned heat dissipation devices are only ∼15 W m−2 K−1.12 It is therefore necessary to continue to develop novel passive heat dissipation devices with better natural convection heat transfer properties.

Carbon nanotubes (CNTs)14 have special structures and excellent thermal properties. The ambient-temperature thermal conductivities of single CNTs could reach up to 3000−5000 W m−1 K−1,15−18 which are obviously higher than that of common metals (e.g., aluminum and copper). So it is believed that CNTs would have promising applications in thermal management devices. Although the thermal conductivity of CNTs have been richly investigated, the heat dissipation properties of CNTs are rarely explored. More importantly, natural convection heat transfer coefficients of suspended rods are known to increase with decreasing diameters (d).19,20 So it is believed that the nanoscale rod as small as individual CNTs would have much higher natural convection heat transfer coefficient than common macroscale materials.20 Furthermore, individual CNTs have very high surface-to-volume ratios, which could significantly enhance the heat dissipation to surrounding air,21 and several previous works have proved that heat transfer coefficients between CNTs and surrounding air are significantly higher than bulk materials.20,22−27 So the CNT, which is a kind of isothermal superconductive materials, could not only maintain homogeneous temperature but also reinforce the thermal convection. Additionally, the radiation heat transfer properties of well-dispersed CNTs, which could give a quantized lattice motion independently and act as “molecular cooling fan”, have been well-studied by the Lin group.28,29 Their results indicated that the well-dispersed CNTs could obviously enhance the radiation heat transfer properties of common metals. However, almost no CNT-based heat dissipation devices for passive electronic cooling applications have been reported. On the one hand, highly oriented 31203

DOI: 10.1021/acsami.6b08815 ACS Appl. Mater. Interfaces 2016, 8, 31202−31211

Research Article

ACS Applied Materials & Interfaces

(ZEISS Imager A1), as shown in the Figure 1f. The micrograph shows that the microchannels were regularly etched on the copper plate and that there is a lot of oxidization layer. So to clean the oxidization layer, these sculptured copper plates were soaked in hydrochloric acid and triple-rinsed in deionized water. Nowadays, the SACNT arrays (Figure 1c) are massively synthesized using the chemical vapor deposition (CVD) by our group. Furthermore, continuous SACNT films can be directly pulled from the SACNT arrays (Figure 1d).30,32,37 Acetylene was used as precursor, and the SACNT arrays were grown on silicon wafers at the temperatures of 650−700 °C. The growth time was in the range of 5−20 min, and the array height was altered from 100 to 900 μm.30,37 Most of the CNTs are multiwalled CNTs, and their diameters range from 10 to 20 nm. There are strong van der Waals forces between neighboring single CNTs in the SACNT arrays.30,37 Then an electrical motor, which had a uniform speed, was used to pull the monolayer SACNT films from the SACNT arrays.30,37,38 The schematic diagram in the Figure 1d shows the fabricating principle of the monolayer SACNT films. The monolayer SACNT film is very thin (only a few dozens of nanometers), and 80% of light can penetrate the films.37,39 All individual CNTs in the SACNT films are high-directionally arranged, which results in the anisotropic properties of SACNT films.37 Specifically, SACNT films have much better thermal and electrical conductivities along the alignment directions than that across the alignment directions of individual CNTs.31 The scanning electron microscope (SEM; Nova NanoSEM) image of lying and suspending monolayer SACNT films is shown in the Figure 1g, which indicates that both individual CNTs and CNT bundles exist in the SACNT films. The SEM image also shows that there are numerous interspaces between individual CNTs in the SACNT films. The lying SACNT films over underlying microchannels were fabricated by the following methods. First of all, as mentioned above, the monolayer SACNT films were pulled from the SACNT arrays by a uniform-speed motor. Second, the etched Cu plates were placed on a lifting platform under the drawn monolayer SACNT films. Meanwhile, the microchannel directions were perpendicular to the drawing direction. Then the platform was lifted slowly until the copper plate contacted with the monolayer SACNT films. Finally, the SACNT films were cut by a cutting laser. The central wavelength and the cutting power of the cutting laser are 1070 nm and 100 W, respectively.37 The spreading direction of SACNT films was guaranteed by the above methods, and a portion of lying SACNT films could be suspended on microchannels. Some ethanol spray was dropped on the samples by an ultrasonic nebulizer. The ultrasonic nebulizer could generate micrometer ethanol drops, which could dropwise add small ethanol drops on the heat dissipation device and would not destroy the suspended SACNT films. After the ethanol drops evaporated, owing to the tensions of ethanol molecules, the SACNT films adhered on microchannel walls. This strategy was used to strengthen the adhesion strength of SACNT films in our previous work.37 Then the adhesion of SACNT films on copper plates was tested according to the ASTM 3359D and ISO 2409 standard. The test results show that the adhesion strength is not very strong because of the relatively weak van der Waals forces between CNTs and copper plates. Figure 2 shows the SEM

macroscopic materials of CNTs, such as superaligned CNT (SACNT) arrays,30 SACNT films,31,32 SACNT fibers,33 and SACNT buckypapers,34,35 have been widely studied. On the other hand, all of these CNT macroscopic materials could have excellent thermal properties.33,36 Additionally, individual CNTs and their macroscopic materials have much lighter weight and more stable chemical properties than common metals. These advantages would be more promising when they are applied in the portable electric devices and deep-space crafts. In our previous work, we have shown that the natural heat dissipation properties of bare copper (Cu) plates were enhanced ∼13% by coating multilayer SACNT films.37 In this paper, a kind of patterned structure of SACNT films and underlying microchannels was designed, and the passive cooling properties were studied. The novel cooling devices based on SACNT films and underlying microchannels have much more effective cooling performances than bare Cu plates (which is widely used as raw materials to fabricate traditional heat dissipation devices). Our results show that enhancements of total natural heat transfer coefficients of these novel cooling devices over bare copper plates range from 58.0% to 86.6%, depending on the heating powers. Meanwhile, the surface emissivity of every sample was measured by a thermal property testing device.37 Then the radiation and natural convection heat transfer properties were respectively calculated. The calculations indicated that the excellent cooling properties of the CNT-based devices are primarily caused by the reinforcement of the natural convection heat transfer properties. Finally, the heat dissipation mechanisms of the CNT-based devices were discussed according to a two-region model. It turns out that enhancements of the total heat dissipation properties are mainly due to the synergistic effects of microchannels and SACNT films, which could obviously reinforce the convection heat transfer properties. Distinctive from previous studies, our results reported here offer insights and an avenue to achieve effective passive heat dissipation devices based on CNTs.



EXPERIMENTAL SECTION

Preparation of the CNT-Based Heat Dissipation Device. The novel CNT-based heat dissipation devices consist of SACNT films and underlying microchannels on copper plates. As shown in Figure 1, microchannels on copper plates were prepared by the following steps. First, three red copper plates (2 cm × 2 cm × 1.5 mm) were polished and rinsed in acetone and deionized water. Then microchannels were sculpted on the copper plates by high-power cutting laser (Beijing Lantian Laser Processing Center), shown in the Figure 1a,b. In the present work, to investigate the correlations among microchannel geometry and thermal performance of the CNT-based heat dissipation device, three kinds of microchannels with different geometries were prepared; detailed information is shown in the Table 1. During the laser sculpturing process, copper plates were oxidized. Morphologies of etched copper plates were characterized by optical microscope

Table 1. Geometry Information of Copper Plates and Etched Microchannels copper plate dimensions samples

width

length (L)

bare copper device 1 device 2 device 3

2 cm

2 cm

1.5 mm

2 cm 2 cm 2 cm

2 cm 2 cm 2 cm

1.5 mm 1.5 mm 1.5 mm

thickness

microchannel dimensions width (w)

depth (d)

interval

200 μm 100 μm 100 μm

500 μm 200 μm 500 μm

100 μm 100 μm 100 μm

Figure 2. SEM images of (a) lying SACNT films covered underlying microchannels and (b) suspended SACNT films on microchannels and contacted interfaces between SACNT films and microchannel walls. 31204

DOI: 10.1021/acsami.6b08815 ACS Appl. Mater. Interfaces 2016, 8, 31202−31211

Research Article

ACS Applied Materials & Interfaces

Figure 3. (a) Schematic diagram of the thermal property testing device. (b) Voltage−current curve of the electric heater, which shows that the resistances almost keep constant during our measurements. images of the device 1 coated by lying SACNT films, which clearly demonstrates that the microchannels are uniformly covered by lying SACNT films, and individual CNTs or CNT bundles have very good alignment (Figure 2a). To show the changes of morphologies, the SEM image of the partly covered etched copper plate by lying SACNT films is shown in the Figure 2b. On the one hand, the SEM image demonstrates that the SACNT films contacted with the microchannel walls and that the lying SACNT films have good suspended conditions. On the other hand, the SEM image also shows that the semiclosed microchannels were generated by the SACNT films. As a result of these microscale porous structures, the air convection was reinforced in microchannels, which could carry out more heat from the microchannel walls and CNTs. Some previously reported papers demonstrated that the convection heat transfer coefficient of CNTs increases with the decreasing diameter.20,26,27 So to take full advantage of individual CNTs or CNT bundles, the monolayer SACNT films were used as coatings of microchannels. Furthermore, interstitial flow of air from open space across SACNT films to microchannels would also reinforce the air convection in the microchannels.40−42 So the superior heat transfer properties and the special structure of the CNTbased heat dissipation device could significantly improve the heat dissipation properties at natural condition. Principle of Heat Dissipation. At the natural condition, the heat could transfer from materials to surroundings via convection, radiation, and conduction. First, the convection heat transfer property is described by the equation

Pc = hA(Th − T0)

Finally, owing to the relatively low thermal conductivity of the air (kair ≈ 0.026 W m−1 K−1),37,45 the conduction heat power is negligible compared with the convection and radiation heat power.37,46 In conclusion, the overall dissipated heat power (P) at natural condition is P = Pr + Pc

Then the natural convection heat transfer coefficients are determined by

h=

−1

2

where Pc(W), h (W m K ), and A (m ) are convection heat power, natural convection heat transfer coefficients, and surface area, respectively. Th and T0 (K) are material temperatures and ambient temperatures, respectively. Note that, for traditional microfins, the surface areas were calculated from the dimensions of fins in the previous works.8,12,13 These calculations led to the facts that convective heat transfer coefficients decrease with increasing microfin heights and that flat plates have higher heat transfer coefficients than microfinned plates.12,13 However, the higher heat transfer coefficient does not imply the higher heat dissipation properties in these conditions.12 Here, for the comparison of the heat transfer coefficients between the bare copper plate and the CNT-based heat dissipation devices, the macroscale apparent area of every sample, which are the same with the bare copper plate, were used as A to determine the natural convection heat transfer coefficients. Second, the radiation heat transfer properties are expressed by the Stefan−Boltzmann equation Pr = σεAFi − k(Th 4 − T0 4)

P − σε AFi − k(Th 4 − T0 4) A(Th − T0)

(4)

In the present paper, the surface emissivity and natural convective heat transfer coefficient were measured by a thermal property testing device, which is shown in the Figure 3a. The Thermal Property Testing Device. In our previous work, a thermal property testing device was reported to measure the heat dissipation properties.37 As is shown in the Figure 3a, the thermal property testing device consists of a flexible electric heater, a porous heat insulator, measured samples (the bare copper plate or the CNTbased heat dissipation devices), an infrared (IR) thermometer, and three k-type thermocouples.37 The electric heater (2 cm × 2 cm × 1.5 mm) consisted of nickel−chromium alloy wires and silicon rubber layers. A KEITHLEY 2400 electric source was used to control the input powers and heating temperatures of the electric heater. The heating powers range from 0 to 5 W, and the corresponding heating temperatures range from ∼300 to 520 K.37 The voltage−current curve of the electric heater is shown in the Figure 3b, which shows that the resistance of the nickel−chromium alloy wires was almost constant during six different heating powers. The porous heat insulator (2 × 2 × 3 cm3) and the electric heater was pasted by a kind of adhesive sponge tape, as well as tested samples were pasted on the another side of the electric heater by high-purity conductive silver paint, which could enhance the thermal interfacial conductance between the heater and tested samples. Hence, we could obtain much better heat conduction properties in the heater-sample side than other sides. Additionally, two k-type thermocouples were set between the copper plate and the heat insulator to simultaneously detect the temperature on samples.37 Another k-type thermocouple was pasted on tested samples to detect the temperature on outer surfaces. An IR thermometer (OPTRIS LS) was set to detect the temperature on the outer surface of samples. The spectral range and temperature range of the IR thermometer are 8−14 μm and 240−1170 K, respectively. The surface emissivity was determined by the synchronous measurement of temperatures by thermocouple and IR thermometer.46 More details about determining methods of the surface emissivity are described in the Supporting Information. At stable states, the overall dissipated heat power (P) was equal to the input power (Pin). Then the power dissipating from the insulator side was calculated and deducted from the total heating power. Additionally, all measurements were operated in a transparent plastic box (1 × 1 × 1 m3). The box was big enough to

(1) −2

(3)

(2)

where Pr(W) is the thermal radiation power, and ε is the surface emissivity. The surface emissivity of materials is significantly influenced by the surface roughness, material type, and surface properties. σ and Fi‑k are the Stefan−Boltzmann factor (5.67 × 10−8 W m−2 K−4) and view factors, respectively.11,43,44 Detailed information and calculating equations of view factors were shown in the Supporting Information. 31205

DOI: 10.1021/acsami.6b08815 ACS Appl. Mater. Interfaces 2016, 8, 31202−31211

Research Article

ACS Applied Materials & Interfaces

Figure 4. (a) Temperature rising curves of the three devices, the etched copper plate, and the bare copper plate when the heat power is 0.42 W. The dimension of etched copper plate is the same with that of the Device 1. (b) The temperature cooling curves of the three devices, the etched copper plate, and the bare copper plate when the heat power is 1.26 W. avoid boundary-layer effect between the testing device and the box.37 In this way, the influence of different airflows on our measurements could be eliminated.

cooling efficiency of the three devices are calculated 22%, 23%, and 25%, respectively. This feature clearly demonstrates that the heat dissipation devices have much more effective cooling properties than the bare copper plate. Additionally, these temperature rising curves in the Figure 4a also show that the CNT-based heat dissipation devices reach to the stable temperatures much faster than the bare copper plate. To be specific, the stable time of the bare copper plate is ∼25 min, while the stable times of the three CNT-based heat dissipation devices are only ∼10 min. Moreover, for the comparison of the heat dissipation properties between the novel CNT-based cooling devices and the etched microchannels (also known as microfins)12 on copper plates, the temperature rising curve of an etched copper plate without SACNT film coatings was also shown in the Figure 4a. The etched copper plate has the same dimension with the Device 1 (Table 1). It turns out that the etched copper plate has lower equilibrium temperature than the bare copper plate, which indicates the etched copper plates with microfins have more effective cooling performances than the bare copper plate. This is mainly because the etched copper plate could extend the heat dissipation area. On the one hand, this strategy has been widely developed to enhance the heat dissipation properties of passive heat cooling devices, and we obtained the similar results with other reported works.12,13 On the other hand, the CNT-based heat dissipation devices reported in this work have much lower equilibrium temperature than the etched copper plate. This characteristic declared that the CNT-based heat dissipation devices have more effective cooling performances than the microfinned copper plates. So we believe, on the one hand, that the heat dissipation devices reported here could replace the traditional metal-finned heat dissipation devices and would be widely used as passive electronic cooling devices. On the other hand, when the temperature reaches steady states at 1.26 W, the input power was turned off, and the tested samples were cooling at the natural codition (Figure 4b). The equilibrium temperatures of the three heat dissipation devices are 83.1, 87.5, and 92.2 °C, respectively. At the same time, the equilibrium temperatures of the etched copper plate and the bare copper plate are 106.2 and 113.3 °C, respectively. Then the cooling efficiency of the three CNT-based devices is calculated ∼19−27%. These features of the temperature curves in the Figure 4a,b clearly show that the CNT-based heat dissipation devices have much higher passive cooling properties than the bare and etched copper plates. Furthermore, weights of the three novel heat dissipation devices and the bare copper



RESULTS AND DISCUSSION Temperature Rising and Cooling Curves. In the present work, six heating powers (in the range of 0.42−1.26 W) were utilized, and corresponding equilibrium temperatures of every sample were recorded in turn. An eight-channel data collector (ART DAM 3039F) with a recording precision of 0.1 K was used to collect temperatures of the three thermocouples,37 and the data collector was connected to a computer by an “RS-485” bus.37 The average value of the temperatures measured by the three thermocouples was adopted. To precisely measure the steady temperatures, the temperature data were continuously recorded for ∼30 min. The temperature rising curves of the three devices, the etched copper plate, and the bare copper plate at the heating power of 0.42 W were shown in the Figure 4a. Other temperautre rising curves at the another five heating powers have the similar features with the curve in the Figure 4a. For clarity of representation, other temperature curves were omitted here. The temperature rising curves shown in the Figure 4a indicate that the equilibrium temperatures of the novel CNTbased heat dissipation devices (i.e., Device 1, Device 2, and Device 3) are much lower than that of the bare Cu plate. Specifically, when the heat power is 0.42 W, the equilibrium temperature of the bare copper plate is ∼57.1 °C. At the same time, the equilibrium temperatures of the Device 1, Device 2, and Device 3 are ∼42.8, 44.5, and 43.9 °C, respectively. For the purpose of excluding the influence of thermal boundary resistances between CNTs and Cu plate on the equilibrium temperatures of these devices, the role of the CNT-copper thermal boundary resistances is briefly discussed here. Several previously reported results have demonstrated that the thermal interfacial conductance between CNTs and solids ranges from 4 to 100 MW m−2 K−1,2,47,48 while the thermal interfacial conductance between CNTs and air are only in the range of 0.075−0.1 MW m−2 K−1.22,24 So it is believed that the CNTcopper thermal boundary resistances could have tiny influence on the heat dissipation properties of these CNT-based cooling devices. Here, the cooling efficiency of the CNT-based heat dissipation devices could be defined as (ΔT/TB), where ΔT is the difference of equilibrium temperatures between the heat dissipation devices and the bare copper plate, and TB is the equilibrium temperature of the bare copper plate. Then the 31206

DOI: 10.1021/acsami.6b08815 ACS Appl. Mater. Interfaces 2016, 8, 31202−31211

Research Article

ACS Applied Materials & Interfaces

Figure 5. Very high heat transfer coefficients between the CNT-based heat dissipation device and surrounding air. (a) Total natural heat transfer coefficients (H) of the device 1, the etched copper plate, and the bare copper plate. (b) Natural convection heat transfer coefficients (h) of the device 1, the etched copper plate, and the bare copper plate.

have advantages of no extra energy consumption, free-noise, and small size. So the novel heat dissipation devices are likely to replace the traditional heat dissipation devices consisting of heat dissipation fins and cooling fans. Furthermore, the natural convection heat transfer coefficients (h, W m−2 K−1) were calculated by deducting the thermal radiation power from the input heating power (shown in the Figure 5b). The thermal radiation power was calculated by the eq 2, and the h was calculated by the eq 4. Here, the surface emissivity of all the tested samples was determined by the synchronous measurement of temperatures using a thermocouple and an IR thermometer.37,46 Eyassu et al.28 have reported that the thermal emissivity of aluminum panel is enhanced from 0.15 to 0.98 by the CNT coatings. However, the monolayer SACNT films are very thin, and the surface emissivity of copper plate is slightly improved. Specifically, the value of the surface emissivity of the bare Cu plate is ∼0.21 ± 0.05, while the average surface emissivity of the three novel heat dissipation devices is ∼0.37 ± 0.08. Because of the oxidation of copper surface, all of the three CNT-based heat dissipation devices have slightly different values of surface emissivity at different heating powers. The thermal radiation is closely related to the surface emissivity and the surface temperatures of tested samples. For bare metallic materials, the surface emissivity is relatively low (